Multivariate characterization of phenotypic traits of Arabia, the main Algerian goat breed

N A Ouchene-Khelifi, N Ouchene, A Da Silva¹ and M Lafri

Science Veterinary Institute, University of Blida, 09000 Blida, Algeria
nakhelifi@gmail.com
¹ Univ. Limoges, INRA, PEREINE EA7500, USC1061 GAMAA, F-87000 Limoges, France

Abstract

In Algeria, goats, in spite of their economic importance, are largely neglected. Goat management is traditional and indiscriminate cross-breeding is current. The phenotypic variability of the main breed, the Arabia (commonly called Arbia by breeders), was investigated for the first time, using multivariate discriminant analysis. The sampling was designed in order to take into account the three ecotypes encountered in the breeding area of the Arabia. A total of 558 females and 133 males were phenotyped in a large area including the cradle of the breed, with 23 quantitative measures and 10 qualitative traits. This study defined finely the breed phenotypically. The informations reported in this study are the first step of the conservation and selection programs for this breed showing an untapped potential with a live weight of 36.6 ± 0.3 kg (fem.) and 47.0 ± 0.9 kg (males) and a withers height of 70.3 ± 0.2 cm (fem.) and 75.1 ± 0.5 cm (males) for example. The Compact Index indicated that Arabia was more suited for meat production and the Foreleg Length index indicated animals with relatively long legs, and hence more adapted to plains and long treks. Our results suggest a large intra breed phenotypic homogeneity in spite of the diversity of the production environments, probably induced by economical exchanges. They are in favor of poor intra-genetic variations for this breed, and indicate that measures have to be taken for the breed preservation.

Key words: Arbia, local breed, analysis, variable

Introduction

Goats are widely adapted to different climates and are found in all production systems. In particular they play important economical role for the small farmers of landless, given their suitability to poor systems (Khan et al 2006).

In Algeria, goats are estimated at 3.8 million of heads, corresponding to 14% of the stock of small ruminants and cattle (MAP 2009). Algeria is only 40% self-sufficient in milk (MADR 2004). Milk production showed significant increase between 1986 and 1996 (from 0.75 to 1.1 million tonnes), but still mainly corresponds to cattle (60%) with very small amounts from sheep (26%) and goats (13%). This low productivity cannot be only explained by harsh climatic conditions of the country but also by lack in the livestock management (Nedjraoui 2006).

In Algeria, goats are distributed unevenly in different regions and under various climatic and environmental conditions (Bourabah et al 2013), principally in the steppic regions (41.1%), in mountainous areas (28.8%) and in Saharan regions (22.5%) (Khemici et al 1993).

Among the four main local breeds (Arabia, Makatia, M’Zabyte and Kabyle) (Ouchene-Khelifi et al 2015), Arabia is the dominant breed with a number of heads estimated at 810,000 (MAP 1998). This breed, related to the Nubian breed (Charlet and Le Jaouen 1977), presents economic interest related to good reproductive performances of fertility and prolificacy (MAP 1998). It is mainly reared for meat; the very limited milk production is merely for domestic consumption. Moreover the breed is able to survive under stressful environmental conditions, including poor nutrition and high temperatures (ANRG 2003).

Phenotypic characterization studies are the first step and the cheapest on the path of the productivity improvement (Odubote 1994). Moreover, according to FAO (2012), this process, including the description of their external and production characteristics in a given environment and under given management practices, is unavoidable for planning the management of the resources.

The phenotypic variability of the four main native breeds (Arabia, Makatia, M’zabite and Kabyle), and of two exotic breeds (Alpine and Saanen), was investigated in Algeria, using multivariate discriminant analysis (Ouchene-Khelifi et al 2015) and suggested cross-breeding practices. To take the analysis one step further the aims of this study is to realize a fine morpho-biometric characterization focusing on the dominant Arabia (Arbia) breed. To this end, the phenotypic variability of the Arabia was examined by considering the three main eco-zones of the breed and its cradle.

Material and methods

Goat sampling

The regions sampled included the cradle of the breed (i.e. limited by the region of Tebessa in the east and the region of Saida in the West) and the areas in which Arabia shows higher concentration (ANGR 2003). Breeders were selected in order to optimize the probability of working with pure Arabia goats, i.e. only individuals certified as being uncrossed Arabia according to the breeder were considered. Thus, 66 villages involved in indigenous Arabia rearing were identified in 17 Algerian regions (Figure 1). The protocol of the phenotypic characterization is described in Ouchene-Khelifi et al (2015) overall 691 goats of Arabia breed were scored comprising 558 females and 133 males. The descriptive characteristics (23 quantitative and 10 qualitative variables) were adapted from the standard cattle breed descriptor list of FAO (2012). All measurements were carried out by the same group of persons, during spring 2013, in order to avoid inter-individual variations.

Figure 1. Geographical location of the sampling of herds’ goats in Algeria, covering the cradle of the Arabia breed (each area marked with a black borderline on the map corresponds to an Algerian wilaya).

Production environment

Our sampling area is divided by two major mountain ranges, the Tell Atlas in the north and the Saharan Atlas in the south. These mountains separate the area of the study into three ecotypes distinguished by their topography, altitude, the nature of the soil, the vegetation diversity and the climatic conditions.

In the north,forests, scrub and matorrals in the Tellien system are the dominant vegetation formations. Rainfall varies between 450 and 1500 mm/year. The average temperature of the coldest month is between 0°C and 9°C and the average temperature of the hottest month varies from 28°C to 31°C (Nedjraoui 2006).

In the middle area, (i.e. located between the Tell Atlas and the Saharan Atlas), the steppe is dominated by Alfa (Stipa tenacissima), Esparto (Lygeum spartum), sagebrush Artemisia ( Artemisia herba alba) representing interesting pastoral value (Nedjraoui and Bedrani 2008). Altitudes vary between 400 and 1200 m. The climate is semi-arid to arid and characterized by low rainfall (100 to 450 mm per year) and high thermal amplitudes (Bneder 2006).

In the south, is the Sahara (80% of the total area country). The sparse vegetation corresponds to hygrophile and psamophile plants highly adapted to xeric requirements. Pastures consist of Graminaceae species of Aristida pungens, Panicum turgidum, , and fodder shrubs such as Acacias. Precipitations are scarce (Nedjraoui 2006).

Statistical analyses

The sexes were analysed separately due to differences in sample size and to the sexual dimorphism. For female goats, the large sample size allowed to realize further statistical analyses and to investigate in which way morpho-biometric characteristics of the Arabia vary or not, according to the eco-type zones.

For quantitative variables: means, standard errors of the mean and coefficients of variation were computed for all of the traits measured. Stepwise discriminant procedure was applied to determine which morphological traits had more discriminant power than others. The relative importance of the morphometric variables in discriminating climatic zones was assessed using the level of significance P<0.01. The multivariate techniques involved the use of principal components and canonical discriminant analysis. The ability of these canonical functions to assign each goat to its group was calculated as the percentage of correct assignment of each group using a linear discriminant analysis procedure.

For qualitative variables: frequencies were computed for all discrete variables considered. Discriminant correspondence analysis was also conducted. The ability of these canonical functions to assign each individual goat to its group was calculated as the percentage of correct assignment of each group.

R R version 3.0.1 (R core team 2012) and Tanagra 1.4.50 (Rakotomalala 2005) were used to conduct all these analyses.

Morphological indexes

Morphological indexes were calculated based on Salako (2006) and Alderson (1999) methods, in order to assess the type and function of the breed. They were calculated as follows:

Weight: body length x girth depth x [(hip width + chest width)/2] /1050

Height slope: wither height – rump height

Length index: body length/wither height

Width slope: hip width/chest width

Depth index: chest depth/wither height

Foreleg length: wither height – chest depth

Balance: (rump length x hip width)/(chest depth x chest width)

Compact index: body weight x 100/wither height

Results

Female quantitative analyses

Females were considered according to their ecotype localization. The sampling was realized as follow: 175 goats were in the middle area, 114 in the north area and 88 in the south area.

The result of the stepwise discriminant analysis showed that 7 of the 23 measured variables were found to be significant (P<0.01). Here we report these seven variables showing best discriminant power according to Wilks Lambda (WL) and F-values: Head Length (WL:0.9, F:20.3), Head Width (WL:0.8, F:15.4), Scapular-Ischial Length (WL:0.7, F:15.4), Cannon Circumference (WL:0.6, F:19.1), Chest Girth (WL: 0.6, F:14.3), Tail Length (WL:0.5, F: 15.1) and Chest Width (WL:0.5, F:6.3). These seven variables are described in detail in the Table 1.

Dunnet contrats showed that north goats were characterized by highest mean values of the Head Length, the Head Width and the Chest Width; the south goats showed the highest mean values of the Chest Girth and the Scapular-Ischial Length; whereas the middle goats were characterized by highest mean values of Tail Length and Cannon Circumference.

The linear discriminant analysis was performed with the 7 variables selected. Results showed that the distribution of the Arabia breed in the three geographic regions could not be considered as confounded groups whereas the high value for the Wilks Lambda has to be underlined (Wilks Lambda=0.5; Bartlett ᵪ²=190, p-value=0.000; Rao F=18.0, p-value=0.000). Two discriminant functions retained 100% of the variance (function 1: eigenvalue=0.41, Rc²=0.5, LW=0.5; function 2: eigenvalue=0.2, Rc ²=0.4, LW=0.8). Figure 2 allows to visualize the linear discriminant analysis. The percentage of individual goats classified into the 3 regions is shown in Table 3. An average of only 64.3% of goats was correctly assigned. Middle and south areas showed best scores (respectively 80.0% and 72.7%). A very bad score was attributed to the north with 40.3% of individuals assigned to north, 41.2% to the middle and 18.4% to the south.

By considering in particular goats from the middle and the south zones, the result of the stepwise discriminant analysis showed that only 4 of the 23 measured variables were found to be significant (P<0.01) (Tail Lenght, Withers Height, Cannon Circumference and Body Weight). Goats from the middle area showed Weight and Withers Height highly significantly inferior to those of the south goats (P<0.001), whereas Tail Length and Cannon Circumference were significantly higher for the goats of the middle area than for the goats of the south one (P<0.001).

Figure 2. Canonical Discriminant Plot; in red goats of the middle area, in blue goats of the southern area and in green goats of the northern area.

Female qualitative traits analyses:

Frequency of each class level for the qualitative traits is showed in table 2. Whatever the criterion group considered, discrete variables tested independently (contingency tables) showed large independence with the factor; all the Cramer’s V test were inferior to 0.2 except for the Coat Color (Cramer’s V test=0.3).

The ability of the canonical functions to assign each goat to its group was calculated as the percentage of correct assignment of each group. Confusion matrix error rate was really high with a value of 62%. Percentage of individual goats classified into the 3 regions for qualitative variables is shown in Table 3. The discriminant analysis showed a clear intra-breed homogeneity taking into account qualitative variables.

Description of males and females according to quantitative, qualitative traits and morphological indexes

For males and females, mean values of morphological variables and their coefficients of variation (CV) are depicted in Table 4. A clear dimorphism appeared, with males showing bigger traits for most measures. Frequencies for discrete variables are computed in Table 4. According to the discrete variables considered, Arabia goats could be described as follow: arc horns were largely present in males and females; hairs were generally black and white, long and for 60% of the sampled animals smooth. Wattles were present in males and infrequently in females; beard and topknot were present in males and only 50% of the females showed such attributes. Ears were medium and mid-drooping in males and in females mostly long and drooping (see photos 1 and 2).

Photo 1. Arabia goat male	Photo 2. Arabia goat female

Morphological indexes are depicted in Table 5. The Arabia breed showed Length Index (LI) and Height Slope (HS) close to 1 in males and females what is characteristic of a short body breed with height at withers and rump quite similar; the Width Slope (WS) was inferior to 1 indicating that width of hip was inferior to chest (particularly for males); the Depth Index (DI=0.4 in males and females) indicated a moderate depth of chest; the Foreleg Length (FL) showed animals high on legs; the Balance (Ba) showed bigger surface for the chest than for the rump; and Compact Index (CI) indicated meat type animals (values above 3.1 for males and females).

Discussion

The Arabia breed was morphologically characterized by the use of 23 quantitative variables and 10 qualitative variables. The females sampled were finely studied through discriminant canonical analysis both for continuous and discrete variables. Considering the qualitative variables, no phenotypic structuration was found according to the 3 ecotype zones sampled. In the same way the quantitative variables showed very limited pattern, with only 7 variables showing distinctions between the ecotypes. Interestingly ecotypes of the middle and the south were differentiated by 4 variables including body weight and wither height; with goats from the middle area showing weight and withers height highly significantly inferior to those of the south goats. Further investigations are required to give an explanation to such results. The ecotype of the north was not discriminated from the two others, which can be explained by the fact that breeders of the steppic and the south areas sell regularly their goats on markets located in the north. Hence goat livestock of the north is mainly constituted of individuals reared in the middle or in the south.

To conclude on this fine scale analysis, the Arabia showed clear morpho-biometric homogeneity in spite of the diversity of its environment productions and in spite of the fact that it has never been subjected to intense selection (the management of the breed is largely traditional). Intra-breed genetic variation is crucial for the resilience capacity of breed submitted to changes in the production and economic environment (Meuwissen 2009) and to respond to artificial selection (Toro et al 2011). Our results postulated in favor of poor intra-genetic variations for this breed, which have to be confirmed by genetic analyses, probably induced by the fact that whatever the production environment considered, commercial exchanges of animals in the markets (largely located in the north) induce high genetic admixture at the land scale. The phenotypic definition of the breed is necessary to the knowledge of this breed (Rothschild 2003). Quantitative measures allowed a fine characterization of the breed. The provisional findings as regards the morphological indexes calculated according to Alderson (1999) and Salako (2006) concerned (i) the Compact Index (3.67 for females); according to Chacon et al (2011) meat type animals have values above 3.15; values close to 2.75 indicate dual purpose and close to 2.60 indicate animals more suitable for milk. Thus Compact Index indicated that Arabia was more suited for meat production. (ii) Moreover Foreleg Length index indicated animals with relatively long legs, and hence more adapted to plains and long treks.

Conclusions

• This study gives the first morpho-biometric characterization of the dominant Algerian goat breed: the Arabia.
  
  
• The Algerian goat livestock is largely neglected and hence unknown.
  
  
• For Algerian breeders, goats rearing is a marginal activity and the preservation of the breed is far from being a preoccupation; that’s why practices of cross-breeding are current.
  
  
• Our results are not very optimistic in term of intra-breed genetic diversity, and indicate that measures have to be taken in favor of the breed preservation.
  
  
• Moreover, this study is a first step for planning the modern management of the resource.

References

Alderson G L H 1999 The development of a system of linear measurements to provide an assessment of type and function of beef cattle. Anim. Gen. Res. Info. 25: 45-55.

ANGR 2003 Rapport National sur les Ressources Génétiques Animales 2003 Bulletin d'information sur les Ressources Génétiques Animales, Algérie. 46p.

BNEDER 2006 Identification et cartographie des zones potentielles à l’agriculture en steppe. Etude diachronique du climat et du bioclimat de la steppe algérienne. Alger, Bureau National d’Etude pour le Développement Rural. pp 47.

Bourabah A, Ayad A, Boukraa L, Hammoudi S M and Benbarek H 2013 Prevalence and Etiology of Subclinical Mastitis in Goats of the Tiaret Region Algeria. Glob. Vet. 5: 604-608.

Chacón E, Macedo F, Velázquez F, Paiva S R, Pineda E and McManus C 2011 Morphological measurements and body indices for Cuban Creole goats and their crossbreds. R. Bras. Zootec. 40: 1671–1679.

Charlet P and Le Jaouen JC 1977 Les populations caprines du bassin méditerranéen Aptitudes et évolution. Options méditerranéennes, 35: 45-53.

FAO 2012 Phenotypic characterization of animal genetic resources. FAO Animal production and health, guidelines. Food and agriculture organization of the United Nations, Rome, pp 142.

Khan H, Muhammad F, Ahmad R, Nawaz G, Zubair R and Zubair M 2006 Relationship of body weight with linear body measurements in goats. J. Agric. Biol. Sc. 13: 51-54.

Khemici E, Mamou M, Lounis A and Bounihi, D 1993 Étude des ressources génétiques caprines de l'Algérie du Nord à l'aide des indices de primarité. Anim. Gen. Res. Info. 17: 61-71.

MADR 2004 Direction Générale des Forêts. Programme d’Action National sur la lutte contre la Désertification. Algérie, Ministère de l’Agriculture et du Développement Rural. pp 104.

MAP 1998 Statistiques des productions animales de l’année 1997. Alger, Ministère de l’Agriculture et de la pêche. 77 p.

MAP 2009 Statistiques des productions animales de l’année 2008. Alger, Ministère de l’Agriculture et de la pêche. 77 p.

Meuwissen T H E 2009 Towards consensus on how to measure neutral genetic diversity? J. Anim. Breed. Genet. 126: 333–334.

Nedjraoui D 2006 Algeria Country Pasture/Forage Resource Profiles. FAO, Rome, pp 28.

Nedjraoui D and Bédrani S 2008 La désertification dans les steppes algériennes, causes, impacts et actions de lutte. Vertigo, la revue électronique en sciences de l'environnement,81.

Odubote I K 1994 Influence of qualitative traits on the performance of West African Dwarf goats. Nigeria J. Anim. Prod. 21: 25-28.

Ouchene-Khelifi N A, Ouchene N, Maftah A, Da Silva A B and Lafri M 2015 Assessing admixture by multivariate analyses of phenotypic differentiation in the Algerian goat livestock. Trop. Anim. Health Pro. 47, 7, 1343–1350. DOI 10.1007/s11250-015-0869-5

R Core Team 2013 R a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria, URL http//wwwR–projectorg/.

Rakotomalala R 2005 TANAGRA un logiciel statistique gratuit pour l’enseignement et la recherche. In Actes de EGC’2005, RNTI-E-3, 2, 697-702.

Rothschild M F 2003 From a sow's ear to a silk purse Real progress in porcine genomics. Cytogenet. Genome Res. 102: 95-99.

Salako A E 2006 Application of morphological indices in the assessment of type and function in sheep. Int. J. Morphol. 24, 13-18.

Toro M A, Meuwissen T H E, Fernandez J, Shaat I and Maki-Tanila A 2011 Assessing the genetic diversity in small farm animal populations. Animal, 5: 1669–1683.

Received 12 March 2018; Accepted 20 April 2018; Published 3 July 2018

Go to top